Intelligent inventory management for automotive spare parts — powered by AI-driven product recognition, real-time stock tracking, and automated bill matching.
StockSync is an inventory management system purpose-built for automotive spare parts dealers. It uses AI-powered image recognition to read product labels, barcodes, and QR codes — converting physical stock into digital inventory entries with minimal manual effort.
Photograph or scan any product label. AI extracts brand, product name, part number, MRP, and barcode data automatically.
Upload purchase invoices first. AI parses line items, then scan each product against the bill — every inventory movement linked to a financial document from the start.
Warehouse staff can send product photos via WhatsApp. Images are processed and added to the inward queue.
Track purchased, sold, and balance inventory across vendors, brands, products, and time periods at a glance.
Automotive parts dealers manage hundreds of SKUs from dozens of brands. Stock movements are tracked through bills and challans in Tally, but the physical inventory on the warehouse shelf is never independently verified against these records. Reconciliation is manual, infrequent, and error-prone — discrepancies between what the books say and what is actually in stock accumulate silently.
Role-based access control ensures each user sees only what they need. Permissions cascade from Admin down.
Business owner / manager
Handles purchase and sale operations
On-floor scanning and verification
Three access surfaces — a mobile app for the warehouse floor, a web dashboard for the office, and WhatsApp for quick image capture.
iOS and Android. Camera scanning, barcode reader, product lookup, and inventory operations. Works offline with sync.
Admin dashboard, detailed reports, invoice upload, bill matching, user management, and full inventory view.
Warehouse staff can send product photos to the business's WhatsApp Business API number. Images are queued, processed by AI, and staged for review.
PostgreSQL database, authentication, file storage, real-time subscriptions, and edge functions for AI processing.
AI Engine: Claude Vision API processes product label images, extracts structured data (brand, product name, part number, MRP, barcode), and parses PDF invoices. Estimated cost per image: INR 1-2.
The core intelligence of StockSync. Products are identified through multiple input methods, with a fallback flow for cases where AI recognition is uncertain.
For each product image, the AI extracts the following fields:
| Field | Example | Required |
|---|---|---|
| Brand | Autoform, Uno Minda, GFI, Tiebond | Yes |
| Product Name | Max Arrow Seat Cover, Alloy Wheel with Hub Cap | Yes |
| Part Number | GLLM-074, W1D205-000M00 | If available |
| Barcode / UID | 1108049828, 25102502385057 | If available |
| MRP | 4200, 7654, 10999 | If available |
| Vehicle Compatibility | Exter 2023, Tata Safari 2021-2025, H-Cross | If available |
| Colour / Variant | Black, Silver/Black, Red/Black | If available |
| HSN Code | 87089900 | If available |
| Manufacturing Date | 09/12/2025, 31-10-2025 | If available |
When the AI cannot confidently identify a product — due to poor image quality, damage, or non-standard labels — the system provides a structured fallback:
Confidence Score: Every AI extraction includes a confidence score (0-100%). Extractions above 85% are auto-accepted. Between 60-85%, the user is shown the result for confirmation. Below 60%, manual selection is triggered.
Products can enter through multiple channels simultaneously. The system prevents double-counting through layered checks before anything is committed to inventory.
The purchase flow uses a bill-first approach: every inward inventory movement is tied to a financial document from the start. A secondary "Quick Inward" mode handles the edge case where goods arrive before the bill.
The standard workflow. The purchase invoice drives the receipt process.
For the edge case where goods arrive before the invoice is available.
The system automatically compares the current invoice price against the last purchase price for each product during bill parsing.
The outward flow mirrors the bill-first approach: every dispatch is tied to a sale bill or challan. A "Quick Sale" mode handles walk-in or counter sales.
The standard workflow. A sale bill drives the dispatch process.
For walk-in customers or counter sales where no formal bill exists upfront.
A unified view of all stock across every dimension — what came in, what went out, and what remains.
All inventory views support the following filter combinations:
| Filter | Description | Applies To |
|---|---|---|
| By Vendor / Supplier | Filter stock purchased from a specific vendor (e.g. Janak Motors) | Purchase, Balance |
| By Client / Buyer | Filter stock sold to a specific client | Sales |
| By Brand | Filter by manufacturer brand (e.g. Autoform, Uno Minda, GFI) | All views |
| By Product / Category | Filter by product type (Seat Covers, Alloy Wheels, Mats, etc.) | All views |
| By Vehicle Model | Filter by compatible vehicle (Exter, Safari, H-Cross, etc.) | All views |
| By Date Range | Filter by purchase date, sale date, or invoice date | All views |
| By Part Number | Search by exact part number or barcode | All views |
| By Invoice | View all items from a specific invoice | Purchase, Sales |
Products approaching out-of-stock are surfaced proactively:
Since the bill-first workflow handles matching at receipt and dispatch time, this section serves as the oversight dashboard — a single view of all invoice statuses, unlinked items, and discrepancies that need attention.
A single screen that gives the business owner complete visibility into operations, stock health, and financial reconciliation.
Total purchased, sold, and balance inventory. Filterable by date range, brand, vendor.
Products below reorder threshold. Sorted by urgency with reorder suggestions.
Recent price changes detected across invoices. Pending approvals highlighted.
All invoices by fulfillment status: fully received, partial, pending bill, pending dispatch.
Timeline of scans, uploads, and system events across all users.
Distribution of inventory by brand and product category.
Proactive alerts ensure nothing slips through. Delivered via in-app notifications and push notifications on mobile.
| Alert | Trigger | Recipients | Channel |
|---|---|---|---|
| Low Stock | Product quantity falls below threshold | Admin, Inventory Manager | Push, In-app |
| Price Change | Invoice price differs from last purchase | Admin | Push, In-app |
| Bill Mismatch | Scanned stock does not match invoice | Admin, Inventory Manager | Push, In-app |
| New Invoice Parsed | PDF upload processed successfully | Uploader | In-app |
| OCR Failure | Product could not be identified by AI | Scanner (staff) | In-app |
| WhatsApp Image Queued | New image received via WhatsApp | Inventory Manager | In-app |
| Dead Stock | Product unsold for configurable period (default 90 days) | Admin | Weekly digest |
| Unlinked Receipt | Goods received via Quick Inward without bill for >3 days | Admin, Inventory Manager | Push, In-app |
| Pending Delivery | Invoice partially received for >7 days | Admin | Push, In-app |
| Pending Dispatch | Sale bill partially dispatched for >3 days | Admin, Inventory Manager | Push, In-app |
| Bill Pending (Sale) | Quick Sale without formal bill for >5 days | Admin | Push, In-app |
Core entities and their relationships in the Supabase PostgreSQL database.
| Table | Purpose | Key Fields |
|---|---|---|
products | Master product catalog | brand, name, part_no, barcode, hsn, category, vehicle_model, mrp, min_stock |
vendors | Supplier directory | name, gstin, address, contact, email |
clients | Customer directory | name, gstin, address, contact |
invoices | Uploaded invoices (purchase & sale) | type, vendor/client_id, invoice_no, date, total, pdf_url, status, fulfillment_status |
invoice_items | Parsed line items from invoices | invoice_id, description, hsn, qty, rate, gst, amount, received_qty, dispatched_qty, item_status |
goods_receipt_notes | Goods Receipt Notes (GRN) | vendor_id, invoice_id, mode, status, notes, created_by, created_at |
dispatch_notes | Dispatch tracking | client_id, invoice_id, mode, status, notes, created_by, created_at |
inventory_entries | Every inward/outward event | product_id, direction, qty, price, source, image_url, user_id, grn_id, dispatch_note_id, invoice_item_id, linkage_status |
price_history | Price tracking per product-vendor | product_id, vendor_id, price, invoice_id, date |
match_results | Bill matching outcomes | invoice_item_id, inventory_entry_id, status, notes |
users | System users | name, email, role, phone, active |
End-to-end flows for the three primary operations.
A phased delivery approach to get core value into production quickly, then layer on advanced features.
Supabase setup, authentication, product catalog, basic CRUD. Flutter app with camera scanning and barcode reader. Claude Vision API integration for label OCR. Manual product selection fallback. Basic inventory list view.
Vendor management. Bill-first purchase inward flow: upload invoice, AI parsing, scan against bill line items, partial delivery tracking. Quick Inward mode for goods-before-bill edge case. GRN creation. Price change detection and alerts. Reconciliation dashboard with invoice status tracking. Web dashboard — inventory views with filters.
Client management. Bill-first sale outward flow: create sale bill, scan for dispatch, partial dispatch tracking. Quick Sale mode for walk-in customers. Dispatch note creation. Low stock alerts and reorder suggestions. Balance inventory views. Sale reconciliation.
Admin dashboard with all widgets. Reports engine with PDF/CSV export. Role-based access control. Activity log and audit trail. Push notifications.
WhatsApp Business API integration for image intake. Stock audit workflow. Dead stock detection. Demand-based reorder predictions. Performance optimization and hardening.